GSTAM: Efficient Graph Distillation with Structural Attention-Matching
This addresses the need for more efficient and accurate graph distillation in graph classification tasks, though it appears incremental as it builds on existing attention-based methods.
The paper tackles the problem of graph distillation for classification by introducing GSTAM, which uses attention maps from GNNs to distill structural information into synthetic graphs, achieving performance improvements of 0.45% to 6.5% over existing methods in extreme condensation ratios.
Graph distillation has emerged as a solution for reducing large graph datasets to smaller, more manageable, and informative ones. Existing methods primarily target node classification, involve computationally intensive processes, and fail to capture the true distribution of the full graph dataset. To address these issues, we introduce Graph Distillation with Structural Attention Matching (GSTAM), a novel method for condensing graph classification datasets. GSTAM leverages the attention maps of GNNs to distill structural information from the original dataset into synthetic graphs. The structural attention-matching mechanism exploits the areas of the input graph that GNNs prioritize for classification, effectively distilling such information into the synthetic graphs and improving overall distillation performance. Comprehensive experiments demonstrate GSTAM's superiority over existing methods, achieving 0.45% to 6.5% better performance in extreme condensation ratios, highlighting its potential use in advancing distillation for graph classification tasks (Code available at https://github.com/arashrasti96/GSTAM).